NO2 and HCHO photolysis frequencies from irradiance

HAL Id: hal-00295685https://hal.archives-ouvertes.fr/hal-00295685

Submitted on 29 Jun 2005

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

NO2 and HCHO photolysis frequencies from irradiancemeasurements in Thessaloniki, Greece

C. Topaloglou, S. Kazadzis, A. F. Bais, M. Blumthaler, B. Schallhart, D. Balis

To cite this version:C. Topaloglou, S. Kazadzis, A. F. Bais, M. Blumthaler, B. Schallhart, et al.. NO2 and HCHOphotolysis frequencies from irradiance measurements in Thessaloniki, Greece. Atmospheric Chemistryand Physics, European Geosciences Union, 2005, 5 (6), pp.1645-1653. �hal-00295685�

https://hal.archives-ouvertes.fr/hal-00295685

https://hal.archives-ouvertes.fr

Atmos. Chem. Phys., 5, 1645–1653, 2005www.atmos-chem-phys.org/acp/5/1645/SRef-ID: 1680-7324/acp/2005-5-1645European Geosciences Union

AtmosphericChemistry

and Physics

NO2 and HCHO photolysis frequencies from irradiancemeasurements in Thessaloniki, Greece

C. Topaloglou1, S. Kazadzis1, A. F. Bais1, M. Blumthaler 2, B. Schallhart2, and D. Balis1

1Laboratory of Atmospheric Physics, Aristotle University of Thessaloniki, Greece2Division for Biomedical Physics, Innsbruck Medical University, Austria

Received: 9 February 2005 – Published in Atmos. Chem. Phys. Discuss.: 15 March 2005Revised: 12 May 2005 – Accepted: 30 May 2005 – Published: 29 June 2005

Abstract. An empirical approach for the retrieval of ni-trogen dioxide (NO2) and formaldehyde (HCHO) photoly-sis frequencies from measurements of global irradiance ispresented in this work. Four months of synchronous mea-surements of actinic flux and global irradiance performed inThessaloniki, Greece by a Bentham spectroradiometer wereused to extract polynomials for the conversion of global ir-radiance to photolysis frequencies [J(NO2) and J(HCHO)].The comparison of these photolysis frequency values to thecorresponding values calculated by spectral actinic flux mea-surements, showed a ratio very close to unity for all J’s witha standard deviation of 12% (2σ) for J(NO2) and 6% (2σ)

for J(HCHO). Additional sets of polynomials were also ex-tracted to allow determination of J(NO2) by spectroradiome-ters with lower upper wavelength limits such as single anddouble Brewer spectroradiometers within acceptable uncer-tainty (corresponding ratio was 1 and standard deviation was12% (2σ) for the method that can be used with double Brew-ers and 20% for the method that can be used for single Brew-ers). The validity of the method under different atmosphericconditions was also examined by applying the polynomialsto another set of actinic flux and global irradiance measure-ments performed in May 2004, in Buchhofen, Germany. Inthis case, comparing J values extracted from the polynomi-als to those calculated from actinic flux, showed equivalentresults, demonstrating that the method can also be applied toother measurement sites.

1 Introduction

The photodissociation of trace gases in the atmosphere is aprocess of great importance for atmospheric chemistry. So-lar ultraviolet radiation drives the photolysis of numerous

Correspondence to:C. Topaloglou([email protected])

important species, such as NO2 and HCHO, contributingto their decomposition and removal from the atmosphere aswell as the formation of highly reactive radicals. Photodisso-ciation is the dominant loss process for NO2 (Eq. 1), produc-ing O(3P) atoms which enable tropospheric ozone formation(Eq. 2). However, NO produced from Reaction (1) is oxi-dized back to NO2, consuming ozone in this case (3) (e.g.He and Charmichael, 1999):

NO2 + hν(λ ≤ 420 nm) −→ O(3P) + NO (1)

O(3P) + O2 + M −→ O3 + M (2)

O3 + NO −→ O2 + NO2 (3)

The NO2 photodissociation process itself as well as theNO-NO2 equilibrium in the NOx, (also influenced by NO2photolysis) are closely related to the radical cycles of OHand HO2 (Kraus et al., 2000) and are an important parameterfor tropospheric ozone concentrations.

Formaldehyde is one of the most important and abundantorganic carbonyl compounds in the atmosphere, observed inboth polluted urban and rural environments. It is mainly pro-duced from fossil fuel combustion and is also a secondaryproduct of photochemical oxidation of methane and other hy-drocarbons (Meller and Moortgat, 2000). The processes R1and R2 presented below achieve the photolysis of formalde-hyde:

HCHO+ hν −→ H + HCO(λ ≤ 330 nm) (4)

HCHO+ hν −→ H2 + CO(λ ≤ 361 nm) (5)

The photolysis of formaldehyde plays a significant role inthe formation of photochemical smog, through the genera-tion of HO2 radicals which participate in O3 catalytic pro-duction (Smith et al., 2002). Both H and HCO producedfrom Eq. (4), yield HO2 radicals as shown in chemical reac-tions presented in Eq. (6) and (7):

H + O2 + M −→ HO2 + M (6)

© 2005 Author(s). This work is licensed under a Creative Commons License.

1646 C. Topaloglou et al.: NO2 and HCHO photolysis frequencies

HCO+ O2 −→ HO2 + CO (7)

It is therefore of considerable importance for atmosphericchemistry to accurately determine the photolysis rates of keyatmospheric compounds such as NO2 and HCHO. In thepast, chemical actinometers and filter radiometers have beenused to measure J(NO2) (e.g. Dickerson et al., 1982; Shet-ter et al., 1992; Lantz et al., 1996; Balis et al., 2002; Shet-ter at al., 2003). Similar measurements for other photolysisrates such as J(HCHO), were not as common as in the case ofNO2 or O3, due to considerable requirements concerning thedevelopment and maintenance of these instruments, so thoserates were either estimated or modeled (Kraus and Hofzuma-haus, 1998). In the last decade, spectroradiometric measure-ments of actinic flux have developed enabling the determina-tion of photolysis rates for several atmosperic species as theintegral:

J=

λ2∫λ1

F(λ)σ (λ, T)ϕ(λ, T )dλ (8)

whereσ(λ, T) is the absorption cross section of the molecule,φ(λ, T) the photodissociation quantum yield, (both depen-dant on wavelength and temperature) and F(λ) the spectralactinic flux.

The actinic flux is defined as the total number of photons(e.g. Madronich, 1987) incident on a spherical surface. Theactinic flux describes the radiation incident on a sphericalsurface such as the molecules of the atmospheric species andis the suitable radiation quantity for photolysis frequenciesdetermination. On the other hand, the global irradiance de-scribes the radiance on a horizontal surface integrated overthe whole upper hemisphere, weighted with the cosine of theincidence angle.

Photolysis rate determination by means of spectroradiom-etry has been reported in the last decade for both NO2 (McEl-roy et al., 1995; Kraus and Hofzumahaus, 1998; Shetterand Muller, 1999; Edwards and Monks, 2003; Shetter etal., 2003) and formaldehyde (Kraus and Hofzumahaus, 1998;Shetter and Muller, 1999). However, since actinic flux mea-surements are not trivial, several approaches have been de-veloped in order to enable the use of global irradiance insteadof actinic flux for photolysis rate determination. Previouswork includes the transformation of global irradiance to ac-tinic flux (Cotte et al., 1997; Kazadzis et al., 2000; Webbet al., 2002; Kylling et al., 2003; Schallhart et al., 2004)which can then be used for photolysis rate determination,as well as empirical methods to derive photolysis rates di-rectly from global irradiance measurements (McKenzie etal., 2002; Kazadzis et al., 2004; Seroji et al., 2004).

This work presents an empirical method to determine pho-tolysis rate values for NO2 and HCHO from measurements ofglobal irradiance from a Bentham spectroradiometer. First,the variability of actinic flux to global ratio as a functionof solar zenith angle (sza) is discussed. Then the empirical

method is presented in detail and is implemented on globalirradiance data. The photolysis frequency values producedby the method are compared to corresponding values derivedfrom measured actinic flux. Additional polynomials are ex-tracted using global irradiance at wavelengths suitable for theuse of Brewer spectroradiometer measurements. The resultscomparing J values derived from this method to those calcu-lated from actinic flux are presented for all cases. Finally, theperformance of the method applied to measurements from adifferent measurement site is examined, showing satisfactoryresults.

2 Data and instrumentation

The UV global irradiance and actinic flux spectra usedin this study to extract the polynomials for the empiricalmethod were measured by a Bentham DTM 300 (Divisionfor Biomedical Physics, Innsbruck Medical University), in-stalled at the roof of the Physics Department in Thessaloniki,at the Laboratory of Atmospheric Physics (LAP) from Marchto July 2003. The instrument performed synchronous mea-surements of actinic flux and global irradiance from 290to 500 nm, in the framework of the INSPECTRO project(Influence of clouds on theSpectral actinic flux in the lowerTroposphere,http://imk-ifu.fzk.de/inspectro/index.html). Inaddition, a double-monochromator Brewer MKIII spectrora-diometer provided UV spectral global irradiance measure-ments in the range of 287–366 nm. Finally, total irradiancefrom a Kipp and Zonnen CM21 pyranometer was used to en-able cloud flagging of the spectra. During the INPSECTROproject the Bentham instrument also measured actinic andglobal irradiance at Buchhofen, Germany, in May 2004.

Detailed information on both the instrument character-istics and calibration details can be found at Kazadzis etal. (2004). A summary of the information on the instruments,the locations and the periods of the measurements that areused in both methods presented in this paper are shown inTable 1.

3 Empirical method for J(NO2) retrieval

3.1 Variability of the ratio of actinic flux to global irradi-ance

The basic idea of this method is the determination of J(NO2)

as a function of solar zenith angle, by the use of global irra-diance and empirical relationships, instead of the direct wayusing actinic flux and Eq. (8). For all J(NO2) calculations inthis work, the absorption cross-section and quantum yield ofDeMore et al. (1997) were used at the temperature of 298◦K.Examining the wavelength dependence of the J(NO2) rate us-ing the molecule functions mentioned above, it appears thatthe spectral weighting of the solar radiation spectrum by thisprocess is largest in the region between 375 and 400 nm. The

Atmos. Chem. Phys., 5, 1645–1653, 2005 www.atmos-chem-phys.org/acp/5/1645/

http://imk-ifu.fzk.de/inspectro/index.html

C. Topaloglou et al.: NO2 and HCHO photolysis frequencies 1647

Table 1. Description of instruments details, location and periods of measurements.

Instrument Measurement Quantity Location Period

Brewer MKIII spectroradiometer Global irradiance 287–366 nm Thessaloniki, Greece 2003Pyranometer Total irradiance (305–800 nm) Thessaloniki, Greece 2003Bentham DTM300 Actinic flux 290–550 nm Thessaloniki, Greece March–July 2003spectroradiometer Global irradiance 290–550 nm

Buchhofen, Germany 13–22 May 2004

FIGURE 1

280 300 320 340 360 380 400 420 440Wavelength [nm]

0.000

0.002

0.004

0.006

0.008

0.010

0.012

F(λ)

*σ(λ

)*φ(λ)

/J(N

O2)

[nm

-1]

0.000

0.020

0.040

0.060

0.080

F(λ)

*σ(λ

)*φ(λ)

/J(H

CH

O) [

nm-1

]

Nitrogen dioxideFormaldehyde, R1Formaldehyde, R2

1

Fig. 1. Relative contribution of each actinic flux wavelength(F(λ).σ (λ).ϕ(λ)) to J(NO2) (right axis) and J(HCHO) (left axis) cal-culation, for sza=24◦, on 12 May 2004.

relative contribution of each wavelength to the J(NO2) cal-culation for 12 May 2003, sza=24◦, is shown in Fig. 1. Thecontribution of the 375–400 region is around 40%. As a re-sult, the integral of global irradiance in this spectral regionwas selected as the global irradiance quantity to be used forthe J(NO2) determination.

Since the method aims to use global irradiance insteadof actinic flux, the variability of downwelling actinic fluxto global irradiance is discussed briefly. The ratio actinicflux/global irradiance as a function of solar zenith angle forwavelength of 390 nm is shown in Fig. 2a. All ratio valuesare above unity since actinic flux is a quantity measured by aspherical detector while global irradiance by a flat one. Low-est values of the ratio (around 1.35) represent small solarzenith angles and low aerosol load, where the highest val-ues (around 2.6) appear around 75 degrees of sza and re-fer to cloudless conditions with small aerosol optical depth(AOD) as well. The overall shape of the ratio distribution isa result of the diminished contribution of the direct compo-nent to the global irradiance as solar zenith angle increases.Consequently, high values of this ratio around 75 degrees,characterize this spectral region because of the more signif-icant contribution of the direct component in comparison tothe diffuse in these wavelengths. Since overcast spectra give

23

FIGURE 2

10 20 30 40 50 60 70 80 90solar zenith angle [deg]

1.20

1.60

2.00

2.40

2.80

actin

ic fl

ux /

glob

al ir

radi

ance 390nm


1.20

1.60

2.00

2.40

2.80

actin

ic fl

ux /

glob

al ir

radi

ance 325nm

Fig. 2. Ratio of actinic flux/global irradiance at 390 nm(a) and at325 nm(b) as a function of sza. All data are from Bentham mea-surements in Thessaloniki.

an almost constant value of actinic/global ratio for all sza,around 1.75, it is apparent that the larger variability of thisratio is to be found at solar zenith angles around 70–75 de-grees. The variabily due to different cloud cover conditionsas well as various AOD loads, account for the vertical distri-bution of the data in Fig. 2a, for a single solar zenith angle.

3.2 Description of the method

J(NO2) values were determined by Eq.8 using spectral ac-tinic flux measurements and the [375–400 nm] global irradi-ance integral (E375−400) calculated from simultaneous globalirradiance spectra. The two quantities are shown in Fig. 3. Asit can be observed, J(NO2) is closely correlated to E375−400and its variability for a given irradiance value is rather small.

The dataset was divided into groups of 5 degree szaand 2nd degree polynomials were calculated from leastsquares fits between the J(NO2) and (E375−400). The poly-nomial coefficients are shown in Table C1 (see Appendix).J(NO2) values were then recalculated using these polynomi-als (Jcalc) and were compared to J(NO2) values from actinicflux spectra using Eq. (8). Their average ratioJcalc

J(NO2)was

www.atmos-chem-phys.org/acp/5/1645/ Atmos. Chem. Phys., 5, 1645–1653, 2005


24

FIGURE 3

0.0 4.0 8.0 12.0 16.0 20.0 24.0E375-400 [W/m2 ]

0.0E+0

2.0E-3

4.0E-3

6.0E-3

8.0E-3

1.0E-2

J(N

O 2

) [s-1

]

solar zenith angle (deg)

15-20

20-25

25-30

30-35

35-40

40-45

45-50

50-55

55-60

60-65

65-70

70-75

75-80

80-85

85-90

Fig. 3. J(NO2) photolysis rates as a function of global irradiance inthe region of 375–400 nm.

1.001±0.124 (2σ) including all solar zenith angles, cloudcover and aerosol load conditions (Fig. 4).

The variability of the ratio shown in Fig. 4 is related tothe representativeness of the polynomials. The scatter fora constant sza is smaller in the region of 40–50 degrees,(2σ=0.12), is consistent with the limited variation of theactinic to global ratio as demonstrated in Fig. 2a. Thismeans that the polynomials can more accurately reproducethe J(NO2) values from the global irradiance value for thesesolar zenith angles. On the other hand, the increase of theratio variability for solar zenith angles larger than 55 degreesreflects the greater variability of the actinic flux to global ir-radiance ratio at these angles (2σ=0.44 for 70–75 degreessza) also shown in Fig. 2a.

In order to decrease the method uncertainty, a parameteri-zation with cloud conditions was examined. The spectra usedin this work were characterized according to cloud coverfrom a methodology proposed by Vasaras et al. (2000). Pyra-nometer data were used to determine the variability of the ra-diation field due to clouds for each scan, which was flaggedas 0 (sun not occluded, stable radiation field), 1 (sun oc-cluded, stable radiation field), or 2 (broken clouds, unstableradiation field). For scans flagged as 0 and 2, data were againdivided into groups of 5◦ sza and new polynomials were ex-tracted for each group to produce J(NO2) from E375−400. Forscans flagged as 1, J(NO2) and E375−400 were very success-fully correlated by a linear approximation, where R2=0.99,so this linear expression was used in this case for all solarzenith angles.

The application of the new – cloud dependent – polynomi-als on the Bentham data, produced an overall mean value of0.999 for the Jcalc

J(NO2)ratio, with a standard deviation of 0.12

(2σ). It is apparent that no significant change is evident forthe entire data set. Small differences between theJcalc

J(NO2)ratio

from the original method to the one with cloud flagging havebeen observed, when examining each set of data (Flag 0, 1and 2) individually, as seen in Table 2. The absolute value of

25

FIGURE 4


0.40

0.60

0.80

1.00

1.20

1.40

1.60

J calc

/J(N

O2

)

mean ratio= 1.001, 2σ = 0.124

Fig. 4. Ratio of J(NO2) calculated by E375−400 and polynomials(Jcalc) over J(NO2) estimated by actinic flux versus solar zenith an-gle. The mean value of the ratio is 1.001 and the standard deviation0.124 (2σ).

J(NO2) derived from Flag 1 spectra is improved by approx-imately 1.5%, using the cloud dependent polynomials. Theremaining flag cases have significantly smaller deviations.

Alternatively, using global irradiance instead of ac-tinic flux in formula (8), a series of pseudo NO2 pho-tolysis frequencies (Jpseudo) were produced and the ratioJ(NO2)/Jpseudo as a function of the global irradiance in-tegral (E375−400) was examined. 3rd degree polynomialsJ(NO2)/Jpseudo=f (E375−400) were retrieved, and their ap-plication showed similar results to those derived from thefirst method described. (Their average ratioJcalc

J(NO2)was

1.002±0.100 (2σ) including all solar zenith angles). Thistechnique is similar to the J(O1D) analysis presented inKazadzis et al. (2004), (using J(O1D)/Jpseudoratio and irra-diance at 325 nm) so the implementation of this method forJ(NO2) data using the integral of the irradiance (E375−400)

and solar zenith angle is also described. The advantage of us-ing the initial method, directly retrieving J(NO2) from globalirradiance, is that input optic spectroradiometers having up-per wavelength limits lower than 420 nm, (e.g. Brewer singleand double spectroradiometers) can be used, as it is describedin Sect. 3.3.

3.3 Alternative polynomials for Brewer spectroradiometers

To enable the use of the method described above by spectro-radiometers that do not reach 400 nm, such as double or sin-gle Brewer monochromators (upper wavelength limit 365 nmand 325 nm respectively), similar approaches were devel-oped by using global irradiance at 360 nm (E360) for doubleBrewer spectrometer and 325 nm (E325) for single Brewerspectrometer, as an independent variable instead of E375−400.A new set of polynomials was extracted for each case, us-ing once more Bentham simultaneous actinic flux and global



Table 2. Mean values of JcalcJ(NO2)

ratio for each group of spectra, using the original polynomials and the cloud dependent polynomials.

Type of spectra Polynomials without cloud flagging Polynomials with cloud flaggingMean ratio±2σ Mean ratio±2σ

Flag 0 0.996±0.076 1.001±0.074Flag 1 1.029±0.160 1.012±0.112Flag 2 1.005±0.144 1.005±0.138

26

FIGURE 5


0.00

0.50

1.00

1.50

2.00

J calc

/J(N

O2

)

Photolysis rates calculated from irradiance at 325nmmean ratio = 1.003, 2σ = 0.214


0.00

0.50

1.00

1.50

2.00

J calc

/J(N

O2

)

Photolysis rates calculated from irradiance at 360nmmean ratio = 0.999, 2σ=0.134

Fig. 5. Ratio of calculated J(NO2) from polynomials by the useof global irradiance at 360 nm to J(NO2) from actinic flux (uppergraph). All data are from the Bentham spectroradiometer. Sameratio for J(NO2) calculated from global irradiance at 325 nm (lowergraph).

irradiance measurements. The corresponding ratios ofJcalcJ(NO2)

for these two cases are shown in Fig. 5.The use of global irradiance at 360 nm as the independent

variable does not significantly change the results of the ap-plication of this method. The ratio ofJcalc

J(NO2)is 0.999 with

as standard deviation of 0.134 (2σ). In the case of 325 nm,this ratio continues to be very close to unity, namely 1.003,however the 2σ standard deviation increases considerably,especially as mentioned before for large solar zenith anglesand reaches 0.214. This means that the average values ofJ(NO2) can still be reproduced, but greater uncertainty is in-troduced by the fact that in the course of the spectral scanthe 325 nm wavelength is measured several minutes beforethe 375–400 nm region, which is highly representative of theJ(NO2) value. Rapid changes of the radiation field causedby moving clouds in front of the sun, could result to a nonrepresentative irradiance value at 325 nm, for this photolysisfrequency calculation.

To test the validity of this method, global irradiance spec-tra from a double Brewer MKIII spectrometer working in

27

FIGURE 6


0.00

0.50

1.00

1.50

2.00

J Bre

wer

/J(N

O2

)

Photolysis rates calculated from irradiance at 360nmmean = 1.004, 2σ = 0.274


0.00

0.50

1.00

1.50

2.00

J Bre

wer

/J(N

O2

)

Photolysis rates calculated by irradiance at 325nmmean = 1.019, 2σ = 0.326

Fig. 6. Ratio of calculated J(NO2) from polynomials by the use ofglobal irradiance from a double Brewer MKIII at 360 nm to J(NO2)

from actinic flux (upper graph). Same ratio for J(NO2) calculatedfrom global irradiance at 325 nm (lower graph).

LAP were used to produce J(NO2) values. The data were cor-rected for calibration standard differences between the twoinstruments and for possible wavelength shifts, using SHI-CRIVM (Slaper et al., 1995), an algorithm that also deconvo-lutes the measured spectra to a standard slit with 1nm FWHM(the methodology is discussed in Kazadzis et al., 2004).The photolysis frequencies produced by Brewer irradiancefrom global spectra and the polynomials, JBREWER, werethen compared to J(NO2) values from Bentham’s actinic fluxspectra, for measurements that deviated from each other lessthan 5 min. The results for the use of both wavelengths(360 nm and 325 nm) from the Brewer double monochroma-tor data are presented in Fig. 6.

The results showed a mean ratio of 1.004±0.274 (2σ) forthe use of E360 and 1.019±0.326 (2σ) for the use of E325 asindependent parameters. The same pattern with the increaseof the ratio variability for large solar zenith angles also ap-pears in these results. The total uncertainty is greater in thiscase, because of additional factors affecting the quality ofthe results, such as measuring differences between the in-struments. Another reason for the dispersion of the ratios



28

FIGURE 7

20 30 40 50 60 70 80 90solar zenith angle [deg]

0.50

0.60

0.70

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

J calc

/J(N

O2

)

Photolysis rated calculated from irradiance at 375-400nmmean = 1.001, 2σ = 0.126 (sza<85 deg)Data from Bentham, Buchhofen, Germany, 13-22 May 2004

Fig. 7. Ratio of calculated J(NO2) from polynomials by the useof E375−400 and Thessaloniki polynomials, to J(NO2) from actinicflux. All data are provided from a Bentham spectroradiometer frommeasurements in 13–22 May 2004 in Buchhofen, Germany.

around unity could be the time difference (the order of a fewminutes) between the Brewer and Bentham scans, a poten-tially significant factor leading to large deviations in caseswith variable cloud cover.

3.4 Implementation of the method on data from differentlocations

Finally, in order to investigate the extent to which the methodis valid for other measurement sites, global irradiance and ac-tinic flux data from the Bentham instrument from a field cam-paign of INSPECTRO in May 2004 were used. The Benthamperformed measurements in Buchhofen, Germany (48.7 N,12.9 E, 340 m) from 13–22 May 2004. J(NO2) values werecalculated from the E375−400 integral and the polynomialsretrieved from Thessaloniki’s dataset. The results were thencompared to J(NO2) directly from actinic flux and Eq. (8).The results are shown in Fig. 7. The meanJcalc

J(NO2)ratio for

sza<85◦ is 1.001 and the 2σ standard deviation is 0.126,showing that the polynomials seem to reproduce the J(NO2)

values satisfactorily. The observation site in Buchhofen is ru-ral, therefore the amount and type of aerosols is expected tobe different compared to the measurements in Thessaloniki.The atmospheric pattern for this period was dominated bypartly cloudy and overcast sky conditions, with occasionalclear sky intervals. The application of the method to thisdata set showed generally good agreement among measuredand calculated photolysis frequencies.

4 Empirical method for J(HCHO) retrieval

A similar empirical method could be used for calculatingHCHO photolysis rates. In the wavelength range of interest

29

FIGURE 8

0.00 0.10 0.20 0.30 0.40 0.50

Irradiance at 325nm (W/m2)

1.40

1.60

1.80

2.00

2.20

J(H

CH

O) /

Jpse

udo

(R

1)

solar zenith angle15-2020-2525-3030-3535-4040-4545-5050-5555-6060-6565-7070-7575-8080-8585-90

Fig. 8. Ratio of J(HCHO)/Jpseudoas a function of irradiance at325 nm.

for the two reactions of HCHO (Fig. 1) the variability of theactinic to global ratio versus the solar zenith angle is smallerthan the one at 390 nm as it is shown in Fig. 2b. However,since the spectral region of interest of the HCHO photolysisis between 280 and 360 nm, special consideration has to begiven to the fact that radiation in this spectral region is af-fected by ozone in the atmosphere. For this reason, insteadof directly associating the HCHO photolysis rates to globalirradiance, the ratio J/Jpseudois used where J is the photol-ysis frequency value calculated from Eq. (8) using actinicflux and Jpseudois the corresponding value when global irra-diance is integrated instead of actinic flux . This ratio is as-sociated with global irradiance at 325 nm (Fig. 8). Both theratio and the irradiance quantity are independent of ozone,so the method is to be used for all total ozone column condi-tions. 3rd degree polynomials are extracted for 5◦ sza groups.To calculate the J(HCHO), the global irradiance at 325 nm isused together with the extracted polynomials and finally aJ/Jpseudoratio is calculated. The ratio is then multiplied withJpseudo integrated from spectral global irradiance measure-ments, to produce the J(HCHO) value. The molecular datafor J(HCHO) determination used in this work are the crosssection from Meller and Moortgat (2000) and the quantumyield from DeMore et al. (1997). The extracted polynomi-als for both chemical reactions are presented in Table C2 andC3.

The method was applied to the measurements performedin Thessaloniki from April to July 2003, to produceJ(HCHO) which were then compared to those calculatedfrom actinic flux. The results are shown in Fig. 9.

The ratio JcalcJ(HCHO)

is practically unity and the 2σ standarddeviation for both chemical reactions is found around 6%.The decrease in the standard deviation value compared to thatof the J(NO2) determination, derives from the fact that thevariation of the actinic flux to global irradiance ratio is con-siderably smaller in the wavelength region of the formalde-hyde photolysis, especially for larger solar zenith angles.



30

FIGURE 9


0.40

0.60

0.80

1.00

1.20

1.40

1.60

J calc

/ J(H

CH

O)

HCHO photolysis-----> H+HCOmean=1.001, 2σ = 0.060


0.40

0.60

0.80

1.00

1.20

1.40

1.60

J calc

/ J(H

CH

O)

HCHO photolysis-----> H2+COmean=1.001, 2σ = 0.068

Fig. 9. Ratio of calculated J(HCHO) from polynomials by the useof global irradiance at 325 nm to J(HCHO) from actinic flux for thetwo chemical reactions of formaldehyde photolysis. All data areprovided from the Bentham spectroradiometer.

When applying this method to Brewer global irradiancedata in order to produce J(HCHO) values, taking into con-sideration all factors already mentioned in the previous para-graph, the mean value ofJBREWER

J(HCHO)equals 0.987 and a 2σ

standard deviation of 0.194 for chemical reaction 1 and0.989±0.186 for Reaction 2. Finally, when applying themethod to the Bentham data in Buchhofen from May 2004,the ratio of the calculated J(HCHO) to those determined fromactinic flux is found to be 0.989±0.076 (2σ) for R1 and0.996±0.084 (2σ) for R2 for sza<85◦ (Fig. 10). This smalldeviation of the mean ratio from unity is a result of the imple-mentation of the polynomials extracted in Thessaloniki, anurban site with several clear sky days during the year, to datafrom a rural and mostly cloudy (at the time of the measure-ments) site. It should also be noted that allσ values presentedin this study using the Bentham instrument data, are relatedto the uncertainty of the method itself and not with uncer-tainties of the cross section and quantum yield functions orinstrumental (measurement) uncertainties. Finally, regardingthe analysis presented including Brewer to Bentham instru-ment ratios, one has to take into account the two instruments’differences. The global spectral irradiance data comparisonfrom the two instruments in the UVB and UVA, showed a ra-tio of 0.971±0.086 (2σ) and 0.973±0.084 respectively (Ben-tham measuring higher) for the reported period.

5 Conclusions

Two approaches of an empirical method for calculatingJ(NO2) and J(HCHO) are presented, based on spectral mea-surements of global irradiance and the use of polynomials,as a function of solar zenith angle. The ratio of J values de-

31

FIGURE 10


0.40

0.60

0.80

1.00

1.20

1.40

1.60

J calc

/J(H

CH

O)

mean = 0.989, 2σ = 0.076 (sza<85 deg)Data from Bentham, Buchhofen, Germany, 13-22 May 2004HCHO photolysis-----> H+HCO


0.40

0.60

0.80

1.00

1.20

1.40

1.60

J calc

/J(H

CH

O)

mean = 0.996, 2σ = 0.084 (sza<85 deg)Data from Bentham, Buchhofen, Germany, 13-22 May 2004HCHO photolysis-----> H2+CO

Fig. 10. Ratio of calculated J(HCHO) for both chemical reactionsfrom polynomials by the use of irradiance at 325 nm and Thessa-loniki polynomials, to J(HCHO) from actinic flux. All data areprovided from a Bentham spectroradiometer from measurements in13–22 May 2004 in Buchhofen, Germany.

rived from these methods to those calculated directly by ac-tinic flux spectra is 1.001 with a standard deviation of around12% (2σ) for J(NO2), and for J(HCHO) 1.001±0.06 (2σ)

for chemical reaction R1 and 1.001±0.06 (2σ) for chemi-cal reaction R2, showing that nitrogen dioxide and formalde-hyde photolysis frequencies can satisfactorily be reproducedby this method within acceptable uncertainty depending onthe data user demands. The method could be best used byscientists studying longer time scale datasets and looking atclimatological conditions and changes. For individual casesone has to take into consideration the statistical analysis andresults that are presented and also the uncertainties resultingfrom instrumental errors.

Alternative polynomials were extracted to enable the useof this method by Brewer spectroradiometers (double andsingle) and the ratio of photolysis rates from a Brewerspectroradiometer to those retrieved from actinic flux mea-surements from a Bentham spectroradiometer show a ra-tio of JBREWER

J(NO2)equal to 1.004±0.274 (2σ) using 360 nm,

1.019±0.326 (2σ) using 325 nm, and for HCHO the ra-tio JBREWER

J(HCHO)is 0.989±0.186 (2σ) for reaction R1 and

0.987±0.194 (2σ) for reaction R2 respectively. The levelof agreement for J(HCHO) is similar to that of the globalirradiance spectra measured by the two instruments for thatspectral region.

Finally the method was applied also to measurements ofthe Bentham spectroradiometer, performed in Buchhofen,Germany, in May 2004 in order to examine whether itcould be implemented to data from another measuring site.The results from comparing the rates calculated by thesemethods to those determined from Bentham’s actinic flux



spectra for sza<85◦, show a ratio of 1.001±0.126 (2σ) forJ(NO2) and for J(HCHO), 0.989±0.076 (2σ) reaction R1and 0.996±0.084 (2σ) for reaction R2, demonstrating thatthe method can be applied to other measurement sites as well.

In the absence of routine actinic flux measurements due tothe special configured optics required, the development of al-ternative methods to retrieve photolysis frequency values forNO2 and HCHO can be very useful for atmospheric chem-istry studies, since photolysis rates are an important input pa-rameter for tropospheric chemistry models. Ultraviolet (UV)global irradiance measurements are very common in mon-itoring stations worldwide and a global network has devel-oped in the last 20 years, providing a long history of qualityassured irradiance measurements. Therefore, the retrieval ofphotolysis rate values from global irradiance measurementsallows the reproduction of extensive time series of photolysisrates for nitrogen dioxide and formaldehyde, within reason-able uncertainty.

Appendix A

Table A1. Polynomial coefficients for the calculation ofJ(NO2)=f(E375−400chemnm). Ci represents the i degree coefficient.

Solar zenith angle C2 (×10−6) C1 (×10−4) C0 (×10−5)

(degrees)

15–20 −9.09 5.53 9.24620–25 −9.64 5.79 −14.54825–30 −7.95 5.37 4.37030–35 −8.34 5.62 −19.15635–40 −7.28 5.47 −10.54140–45 −6.68 5.49 −14.00345–50 −3.97 5.25 −10.26350–55 0.012 5.02 −9.15155–60 3.22 5.02 −10.21660–65 6.32 5.11 −12.00065–70 9.31 5.22 −11.90670–75 24.2 4.66 −4.89575–80 53.5 3.79 3.42280–85 40.8 4.35 0.57885–90 43.6 4.35 0.374

Table A2. Polynomial coefficients for the calcula-tion of J(HCHO)/Jpseudo=f(E325) for chemical reactionHCHO→H+HCO. Ci represents thei degree coefficient.

Solar zenith angle C3 C2 C1 C0(degrees)

15–20 4.283 −4.404 0.193 1.89920–25 6.552 −7.803 1.885 1.63925–30 5.825 −6.070 0.949 1.77830–35 −2.415 0.365 −0.396 1.82935–40 −7.747 4.491 −1.241 1.87640–45 −5.227 1.801 −0.226 1.78545–50 −31.901 16.080 −2.086 1.83350–55 −92.722 41.708 −4.662 1.88955–60 −255.563 95.966 −9.138 1.99060–65 −682.879 202.947 −15.857 2.12065–70 −1029.41 232.180 −12.427 1.94570–75 −863.652 166.211 −5.395 1.81575–80 −5.177.15 626.845 −18.422 1.94180–85 −10842.9 630.335 −6.472 1.79485–90 −76310 1924.930 −7.929 1.790

Table A3. Polynomial coefficients for the calculation ofJ(HCHO)/Jpseudo=f(E325) for chemical reaction HCHO→H2+CO.Ci represents thei degree coefficient.

Solar zenith angle C3 C2 C1 C0(degrees)

15–20 5.202 −5.229 0.366 1.89220–25 6.905 −8.116 1.890 1.64925–30 6.029 −6.383 1.021 1.77630–35 −2.016 −0.160 −0.257 1.82235–40 −7.143 3.741 −1.050 1.86840–45 −6.598 2.211 −0.234 1.78745–50 −32.717 16.183 −2.034 1.83350–55 −100.082 44.429 −4.867 1.89555–60 −295.819 109.447 −10.297 2.02260–65 −797.964 233.588 −17.891 2.15965–70 −1228.350 272.479 −14.237 1.97070–75 −991.823 186.025 −5.420 1.81375–80 −5318.360 650.78 −18.192 1.94180–85 −8042.850 458.765 −1.792 1.77585–90 −135796 3079.43 −11.693 1.799



Acknowledgements.Part of the work was conducted in the frame-work of the INSPECTRO EVK2-2001-00135 project, funded bythe European Commission.

Edited by: A. Hofzumahaus

References

Balis, D. S, Zerefos, C. S, Kourtidis, K., Bais, A. F., Hofzumahaus,A., Kraus, A., Schmitt, R., Blumthaler, M., and Gobbi, G. P.:Measurements and modeling of photolysis rates during the Pho-tochemical Activity and Ultraviolet Radiation (PAUR) II cam-paign, J. Geophys. Res., 107(D18), 8138–8149, 2002, Geophys.Res., 108(D16), 4506–4515, 2003.

Cotte, H., Devaux, C., and Carlier, P.: Transformation of irradiancemeasurements into spectral actinic flux for photolysis rates de-termination, J. Atmos. Chem., 26, 1–28, 1997.

De More, W., Sander, S., Golden, D., Hampson, R., Kurylo, M.,Howard, C., Ravishankara, A., Kolb, C., and Molina, M.: Chem-ical kinetics and photochemical data for use in stratosphericmod-eling, JPL Publ., 97-4, 1–278, 1997.

Dickerson, R. R., Stedman, D. H., and Delany, A. C.: Direct mea-surements of ozone and nitrogen dioxide photolysis rates in thetroposphere, J. Geophys. Res., 87, 4933–4946, 1982.

Edwards, G. D. and Monks, P. S.: Performance of a single-monochromator diode array spectroradiometer for the determi-nation of actinic flux and atmospheric photolysis frequencies, J.Geophys. Res., 108(D16), 8546–8558, 2003

He, S. and Charmichael, G. R.: Sensitivity of photolysis rates andozone production in the troposphere to aerosol properties, J. Geo-phys. Res., 104(D21), 26 307–26 324, 1999.

Kazadzis, S., Bais, A. F., Balis, D., Zerefos, C. S., and BlumthalerM.: Retrieval of down-welling UV actinic flux density spectrafrom spectral measurements of global and direct solar UV irradi-ance, J. Geophys. Res., 105, 4857–4864, 2000.

Kazadzis, S., Topaloglou, C., Bais, A. F., Blumthaler, M., Balis, D.,Kazantzidis, A., and Schallhart, B.: Actinic flux and O1D pho-tolysis frequencies retrieved from spectral of irradiance at Thes-saloniki, Greece, Atmos. Chem. Phys., 4, 2215–2226, 2004,SRef-ID: 1680-7324/acp/2004-4-2215.

Kraus, A. and Hofzumahaus, A.: Field measurements of atmo-spheric photolysis frequencies for O3, NO2, HCHO, CH3CHO,H2O2 and HONO by UV spectroradiometry, J. Atmos. Chem.,31, 161–180, 1998.

Kylling, A., Webb, A. R., Bais, A. F., Blumthaler, M., Schmitt, R.,Thiel, S., Kazantzidis, A., Kift, R., Misslebeck, M., Schallhart,B., Schreder, J., Topaloglou, C., Kazadzis, S., and Rimmer, J.:Actinic flux determination from measurements of irradiance, J.Geophys. Res., 108(D16), 4506–4515, 2003.

Lantz, K. O, Shetter, R. E., Cantrell, C. A., Flocke, S. J., Calvert, J.G., and Madronich, S.: Theoretical, actinometric and radiometricdeterminations of the photolysis rate coefficient of NO2 duringthe Manua Loa Observatory Photochemistry Experiment 2., J.Geophys. Res., 101, 14 613–14 629, 1996.

Madronich, S.: Photodissociation in the Atmosphere: 1. ActinicFlux and the Effects of Ground Reflections and Clouds, J. Geo-phys. Res., 92, D8, 9740–9752, 1987.

McElroy, C. T., Midwinter, C., Barton, D. V., and Hall, R. B.: Acomparison of J-values from the composition and photodissocia-tive flux measurements with model calculations, Geophys. Res.Lett., 22 (11), 1365–1368, 1995.

McKenzie, R. L., Johnston, P. V., Hofzumahaus, A., Kraus, A.,Madronich, S., Cantrell, C., Calvert, J., and Shetter, R.: Rela-tionship between photolysis frequencies derived from spectro-scopic measurements of actinic fluxes and irradiances during theIPMMI campaign, J. Geophys. Res., 107(D5), 4042–4057, 2002.

Meller, R. and Moortgat G. K.: Temperature dependence of the ab-sorption cross sections of formaldehyde between 223 and 323 Kin the wavelength range 225–375 nm, J. Geophys. Res, 105 (D6),7089–7101, 2000.

Schallhart, B., Huber, M., and Blumthaler, M.: Semi-empiricalmethod for the conversion of spectral UV global irradiance datainto actinic flux, Atmos. Env., 38, 4341–4346, 2004.

Seroji, A. R., Webb, A. R., and Coe, H.: Derivation andvalidation of photolysis rates of O3, NO2 and CH2O froma GUV-541 radiometer, J. Geophys. Res., 109, D21307,doi:10.1029/2004JD004674, 2004.

Shetter, R. E., McDaniel, A. H., Cantrell, C. A., Madronich, S.,and Calvert J. G.: Actinometer and Eppley radiometer measure-ments of the NO2 photolysis rate during the Manua Loa Observa-tory Photochemistry Experiment, J. Geophys. Res., 97, 10 349–10 359, 1992.

Shetter, R. E. and Muller, M.: Photolysis frequency measurementsusing actinic flux spectroradiometry during the PEM-Tropicsmission: Instrument description and some results, J. Geophys.Res., 104, 5647–5661, 1999.

Shetter, R. E. Junkermann, W., Swartz, W. H., Frost G. J., Craw-ford, J. H., Lefer, B. L., Barrick, J. D., Hall, S. R., Hofzuma-haus, A., Bais, A., Calvert, J. G., Cantrell, C. A., Madronich, S.,Muller, M., Kraus, A., Monks, P. S., Edwards, G. D., McKenzie,R., Johnston, P., Schmitt, R., Griffioen, E., Krol, M., Kylling, A.,Dickerson, R. R., Lloyd, S. L., Martin, T., Gardiner, B., Mayer,B., Pfister, G., Roth, E. P., Koepke, P., Ruggaber, A., Schwan-der, H., and Van Weele M: Photolysis frequency of NO2: Mea-surement and modeling during the International Photolysis Fre-quency Measurement and Modeling Intercomparison (IPMMI),J. Geophys. Res., 108 (D16), 8544–8558, 2003.

Smith, G. D., Molina, L. T., and Molina, M. J.: Measurement ofradical quantum yields from formaldehyde photolysis between269 and 339 nm, J. Phys. Chem., 106, 1233–1240, 2002

Vasaras, A., Bais, A. F, Feister, U., and Zerefos, C. S.: Comparisonof two methods for cloud flagging of spectral UV measurements,Atmos. Res., 57/1, 31–42, 2001.

Webb, A. R., Kift, R., Thiel, S., and Blumthaler, M.: An empir-ical method for the conversion of spectral UV irradiance mea-surements to actinic flux data, Atmos. Environ., 36, 4044–4397,2002.

Wiegand, A. N. and Bofinger N. D.: Review of empirical methodsfor the calculation of the diurnal NO2 photolysis rate coefficient,Atmos. Environ., 34 , 99–108, 2000.


http://direct.sref.org/1680-7324/acp/2004-4-2215

Documents

NO2 and HCHO photolysis frequencies from irradiance